12342
J. Am. Chem. SOC.1995,117, 12342-12343
Sparteine-Mediated a-Lithiation of N-BOC-N-Methylbenzylamine: Rapid Racemization and Subsequent Deracemization Manfred Schlosser* and Dominique Limat
Scheme 1
spartelnr
1
Li-CHlCHaIC,&
Li-CHICHalC&
4
TMEDA
lnstitut de Chimie organique de 1'Universite Batiment de Chimie CH-1015 Lausanne-Dorigny, Switzerland Received May 17, 1995 More than a quarter of a century ago, the Nozaki group' studied the asymmetric deprotonation of ethylbenzene using butyllithium in the presence of (-)-sparteine to isolate, after carboxylation, I-phenylpropanoicacid (515%) with a maximum enantiomeric excess (ee) of 30%. Sporadically, other investigations of sparteine as a chiral auxiliary were reported? but none of them proved to have a potential for practical development. The breakthrough was made by Hoppe et aL3 who found most impressive enantioselectivities when submitting 0-alkyl, 0-allyl, and 0-benzyl carbamates to a sec-butyllithium/(-)-sparteinepromoted a-hydrogeflithium exchange followed by electrophilic trapping. Shortly afterward, Beak et ale4successfully extended this method to the asymmetric deprotonation of N-tertbutoxycarbonyl-protected pyrrolidines. Most of the organometallic intermediates involved were configurationally stable5 and must hence have been endowed with their nonracemic individuality in the very moment of their appearance. There are, however, a few notable exceptions: when converted into their a-lithiated derivatives in the absence of any enantiodiscriminating reagent and subsequently treated with sparteine before being trapped by an electrophile, (E)-but-Zenyl NJV-diisopropylcarbamate3and N-methyl-3-phenylpr~panamide~ gave products with high (280%) ee values. Obviously, the chiral auxiliary had preferentially intercepted one component from the equilibrium of rapidly interconvertingorganometallic antipodes, thus favoring the formation of one of two possible diastereomeric complexes. We wish to elaborate on this point. As our results demonstrate, benzylic carbamates (la,b) can be a-lithiated in the presence of (-)-sparteine enantioselectively; they then racemize instantaneously, but, after a while, the original homochirality is restored. Furthermore, our data reveal an unprecendented solvent effect on the stereochemical outcome of the electrophilic substitution. (1) Nozaki, H.; Aratani, T.; Toraya, T.; Noyori, R. Tetrahedron 1971, 27, 905-913. (2) Kretchmer, R. A. J. Org. Chem. 1972,37, 2744-2747. Guettb, M. S.; Capillon, J.; Guettb, J. P. Tetrahedron 1973,29, 3659-3667. Okamoto, Y . ;Suzuki, K.; Kitayama, T.; Yuki, H.; Kageyama, H.; Miki, K.; Tanaka, N.; Kasai, N. J. Am. Chem. SOC. 1982,104,4618-4624. (3) Hoppe, D.; Zschage, 0. Angew. Chem. 1989,101, 67-68; Angew. Chem., Int. Ed. Engl. 1989,28, 65-66. Reviews: Hoppe, D.; Kriimer, T.; Schwark, J. R.; Zschage, 0.Pure Appl. Chem. 1990,62,1999-2006. Hoppe, D.; Hintze, F.; Tebben, P.; Paetow, M.; Ahrens, H.; Schwerdtfeger, J.; Sommerfeld, P.; Haller, J.; Guamieri, W.; Kolczewski, S.; Hense, T.; Hoppe, I. Pure Appl. Chem. 1994,66, 1479- 1486. See also: Kaiser, B.; Hoppe, D. Angew. Chem. 1995,107,344-346; Angew. Chem., Int. Ed. Engl. 1995, 34,323-325. (4) Kerrick, S. T.; Beak, P. J. Am. Chem. SOC. 1991, 113, 9708-9710. Beak, P.; Kerrick, S. T.; Wu, S.-d.; Chu, J.-x. J. Am. Chem. SOC. 1994, 116, 3231-3239. In the meantime, this group has also studied the enantioselective metalation and trapping of N-BOC-N-methylbenzylamine (personal communication by Prof. P. Beak, June 1994). See also: Voyer, N.; Roby, J. Tetrahedron Lett. 1995,36, 6627-6630. (5) Recent accounts of the configurational mobility of a-oxy- and a-amino-substituted alkyllithium compounds: Gawky, R. E.; Zhang, Q.h. Tetrahedron 1994, 50, 6077-6088. Elworthy, T. R.; Meyers, A. I. Tetrahedron 1994,50, 6089-6096. Carstens, A.; Hoppe, D. Tetrahedron 1994,50, 6097-6108. (6)Beak. P.; Du, H. J. Am. Chem. SOC. 1993, 115, 2516-2518. See also: Thayumanavan, S.; Lee, S.; Liu, C.; Beak, P. J. A m Chem SOC.1994, 116, 9755-9756. Basu, A,; Beak, P. J. Am. Chem. SOC., submitted.
i I - - - .
sparteine lfull structure1
CH3 HEX or DEE
1
K
H
2& N :COOH 2b: E/ = CSSCH, 2C: N : CH3 l& N = D
F
N
Racemic a-~-N-B:OC-N-methylbenzylamine~ (lb) was vigorously stirred with a solution of sec-butyllithium and (-)sparteine (1 equiv of each) in neat tetrahydrofuran (THF) for 4 h at -75 "C before being poured onto dry ice. In addition to 53% of 1-(N-tert-butoxycarbonyl-N-methylamino)1-phenylacetic acid 2a (El = COOH, H* = 2H, isotopic label 195%), 39% of the starting material was isolated, in which the (S)enantiomer was preponderant to the extent of 25% ee.8 The efficacity of the kinetic resolution improved to 88 and 85% ee when the sparteine-mediated metalation reaction was conducted in diethyl ether (DEE) or hexane (HEX), respectively. The 88% ee (5')carbamate l a thus obtained was metalated with sec-butyllithium in the presence of N,N,hi"-tetramethylethylenediamine in neat THF, DEE, and HEX. Carboxylation after 2 h at -75 "C gave a totally racemic product in all three cases (yields of acid 2a, El = COOH, H* = 2H, 47-63%). No metalation of (S)-lb occurred when sec-butyllithium was used in the presence of sparteine. However, when (5')-lb was consecutively treated with sec-butyllithiumin THF (1 equiv, 4 h at -75 "C), (-)-sparteine (1 equiv, 2 h at -75 "C), and dry ice, optically active acid was formed (82% ee). The deprotonation-promotingtransition state (7) Prepared from [a-2H]benzaldehyde [Chancellor, T.; Quill, M.; Bergbreiter, D. E.; Newcomb, M. J. Org. Chem. 1978,43,1245-12461 by reductive amination [Schlosser, M.; Simig, G. J. Chem. SOC.,Perkin Trans. I 1992,1613-1616; 1993,1631 and subsequent acylation with di-tert-butyl carbonate [Ponnusamy, E.; Fotadar, U.; Spisni, A.; Fiat, D. Synthesis 1986, 48-49]. (8) The enantiomeric composition of lb was determined by 400 MHz 'H-NMR in the presence of 0.1 equiv of Eu(hfc)3 [(R)6 5.13, (S)6 5.06; 0.1 M in CDC13; authentic material by formylation and reduction of (R)[a-2H]benzylamine,which was prepared according to the method of Midland et ai.: Midland, M. M.; Greer, S.; Tramontano, A.; Zderic, S. A. J. Am. Chem. SOC. 1979, 101, 2352-23551. Also determined were 2a in the presence of 1.0 equiv of cinchonidine [(R)6 2.72, (S) 6 2.70; 0.2 M in CDCl3; authentic material: Kajiyama, S.; Irie, K.; Kido, T.; Koshimizu, K.; Hayashi, H.; Arai, M. Tetrahedron 1991, 47, 5433-54621, 2b,after hydrolysis to 2a (0.6 M KOH in CH3OH, 30 min at 65 "C; 95%), and 2c, after hydrolytic removal of the BOC group (1 M HCl in 50% aqueous dioxane, 1 h at 25 "C; 85%) in the presence of 0.05 equiv of Eu(hfc)3 [ ( R ) 6 5.22, (S) 6 5.15; 0.1 M in CDCl3; authentic material: Chantrapromma, K.; Ollis, W. D.; Sutherland, I. 0. J. Chem. SOC., Perkin Trans. I 1983, 1049- 10611.
0002-7863/95/1517-12342$09.00/0 0 1995 American Chemical Society
Communications to the Editor f
100
[%I 80
J. Am. Chem. Soc., Vol. 117, No. 49, 1995 12343 Table 1. Sparteine-Mediated a-Lithiation of N-BOC-N-methylbenzylamine (la) and Enantioselectivities (ee, %) as a Function of the Solvent and the Electrophile El-X (Yields of Products, 43-65%) El-X HEX DEE THF
60
LO
20
Scheme 2 0
Figure 1. Sparteine-mediated a-lithiation of N-BOC-N-methyl-benzylamine (la) and subsequent carboxylation in HEX and THF. Yields and enantioselectivities as a function of the metalation time.
obviously collapses to produce an incompletely coordinated, immediately racemizing species (such as 3) rather than to generate directly the organolithiuddiamine adduct (e.g., 4) (Scheme 1). All the following experiments were carried out with unlabeled starting material l a (H* = IH). Carboxylation of the lithiated intermediate was found to take place in THF, in a 1:l mixture of THFDEE and ethylene glycol dimethyl ether/HEX with inversion9 (85, 26, and 70% ee, respectively), but in tert-butyl methyl ether, DEE, and in HEX with retentiong (61, 67, and 8 1 % ee, respectively). In dimethyl ether, completely racemic acid was formed. The evolution of the reaction has been followed by withdrawing and analyzing samples over a period of 6 h. The increase of yields with time was practically solvent-invariant, reaching the maximum after roughly 1 h (see Figure 1). In HEX and DEE, the ee values started low but climbed rapidly to exceed the 50% level after only a few minutes. In contrast, in THF, little enantioselectivity was observed even after 30 min, although by then metalation was nearly complete. After roughly 40 min, precipitation set in, and simultaneously the ee numbers rose steeply (see Figure 1). The solvent effect on the stereochemistry holds for various kinds of electrophiles, as exemplified with carbon disulfide, methyl iodide, dimethyl sulfate, and o-deuteriophenylacetylene. The methyl dithiocarboxylate 2b (from the lithium dithiocarboxylate and methyl iodide), the methyl homologue 2c, and the deuterated starting material l b had the configuration coresponding to inversion when produced in THF, while the retention mode was obeyed in DEE or HEX. The enantioselectivities observed did not vary significantly with the nature of the electrophile (Table 1).8 To explain the solvent dependence of our reactions, we postulate a dynamic equilibrium to exist between the organolithiudsparteine contact species 4-a, the prevailing ground state structure, and an ion pair 44p, in which the benzyl moiety has been displaced by a solvent molecule (Sv) from its coordination to the metal (Scheme 2). The fraction of ion pairs may remain undetectably small under all circumstances, but it will increase when a good donor solvent such as THF is employed. Beyond a certain threshold value, the electrophilic substitution may pass (9) Retention and inversion with respect to the proton removed. It is a plausible though unproven conjecture that the benzylic center of the organolithium/(-)-sparteine complex has the (S)configuration.
H 4-C.8.
'
CH3
o,::o
4-1.p.
partially, if not exclusively, through the ion pair which should prefer the inversion mode in order to avoid steric crowding on the face shielded by the bulky counterion. In contrast, the contact species should be attacked by the electrophile with retention, as is common. Although compatible with all results collected so far, our mechanistic rationale may be an oversimplification. The crucial organolithiudsparteine intermediate 4 has been modeled according to an X-ray structurelo of a very similar adduct, but carrying an N-pivaloyl instead of the N-BOC moiety. We can nevertheless not rule out the formation of mixed aggregates in which one molecule of the aminobenzyllithium intermediate (1) would coordinate not only with sparteine but also with one molecule of the metalation reagent (e.g., sec-butyllithium). The yields conspicuously centering around 50% (much sec-butyllithium remaining unconsumed) are suggestive in this respect. The solvent control over the stereochemical outcome of sparteine-mediated electrophilic substitution reactions is particularly welcome since the auxiliary is available in only one enantiomeric form. Applications to the synthesis of unnatural amino acids and isoquinoline alkaloids are envisaged. For example, we have deprotected (R)-and (8-N-BOC-N-methyl1-phenylethylamine (75 and 85% ee), prepared as described above, to carry out an N-alkylation with 2,2-diethoxyethyl bromide, followed by ortho-formylation (by consecutive treatment of the intermediate with butyllithium and N,N-dimethylformamide), acid-catalyzed cyclization, and reduction (with potassium tri-sec-butylborohydride), to provide (R)- and (8(1,2-dihydro-l,2-dimethyl-3-isoquinolyl)methanol in 10% overall yield. Independent of these practical considerations, we are intrigued by the emerging possibilities to gain insight into subtle details of organometallic reactions. Acknowledgment. This work was financially supported by the Schweizerische Nationalfonds zur Forderung der wissenschaftlichen Forschung, Bem, Switzerland (Grants 20-36'385-92 and 21-41'22 1.94-Chiral-2). JA95 1603R (10) Boche, G.; Marsch, M.; Harbach, J.; Harms, K.; Ledig, B.; Schubert,
F.;Lohrenz, J. C. W.; Ahlbrecht, H. Chem. Ber. 1993, 126, 1887-1894.
